Barry MacDonald is Director and Rob Walker Senior Research. Associate of the Ford Project on Success and Failure and Recent. ⢠innovation (Ford SAFARI) at ...
:
••.•
•
•••••• • •
•
•
•
•
.
•
:••••••••,•
••••-
•••
: •
•
•
• •••:-:
r
Case-study and the Social Philosophy of Educational Research CARE
••••'•
•••• •
•
.
.
•
.
. •-
. - • . •' .
. • .
. •
• . ••.
.
•
• •
•
;
-
.•
• ••• ••••••,..-•
•••.:
•'•:.•••
.
•
•
•
•
• • .•
-
. ..
.
University of East Anglia
BARRY MACDONALD AND ROB WALKER
-:•••• ..••• •
Reference Only
The methodological difficulties faced by curriculum evaluators who want to offer a comprehensive range of information about new programmes have drawn them to the case-study as a technique. Many of the quite legitimate questions that are put to evaluators, especially by teachers, cannot be answered by the experimental methods and numerical analyses that constitute the . instrumental repertoire of conventional educational research. Such questions are directed at the experience of the participants, and at the nature and variety of transactions which characterise the learning milieu of the programme. There seems to be a need to find ways of portraying this experience and this milieu so that • prospective users of new programmes can relate them to their own experience, circumstances, concerns and preferences. The • case-study is one such way, and evaluators have been prominent among those who are beginning to advocate, and explore in practice, its fruitfulness in educational enquiry. The purpose of this brief paper is to examine the pedigree of the case-study as a research method, to draw attention s to features of our educational system which pose problems for the conduct of case-study, and to suggest some guidelines for its use in education. Case-study is the examination of an instance in action. The choice of the word 'instance' is significant in this definition, because it implies a goal of generalisation. We might say that case. study is that form of research where n=1, only that would be misleading, because the case-study method lies outside the • discourse of mathematical experimentalism that has dominated - - Anglo-American educational research. It is difficult to account for the neglect of the case-study as a general method of educational science, especially in view of its Barry MacDonald is Director and Rob Walker Senior Research Associate of the Ford Project on Success and Failure and Recent innovation (Ford SAFARI) at C.A.R.E.
•
and in Hamilton et al (eds) Beyond the Numbers Game, MacMillan 1977, and in The Urban Review, New York, 1977, and in Case Study Meth ERM881 (Deakin University) and Case Study in Education: ED826 (Univers of Queensland), Deakin University Press 1982. • :.
,
• ;:...
.
-
.•
;
:
•
-••• • •••••
•••:-
'••:;• ' •
•
•••
•
. -
;
; •
'
•
.
CASE-STUDY AND THE SOCIAL PHILOSOPHY OF EDUCATIONAL RESEARCH
Carry MacDonald and Rob Walker
Ford SAFARI Project Centre for Applied Research in Education University of East Anglia
The methodological difficulties faced by curriculum evaluators who want to offer a comprehensive range of information about new programmes have qrawn them to the case•study as a technique.' Many of the quite legitimate questionS that are put to evaluators, especiallytby teachers, cannot be answered by the experimental methodivand numerical analyses that constitute the instrumental repertoire of conventional educational research. Such questions are directed at the. experience of the participants, and at the nature and variety of transeCtions which characterise the learning milieu of the programme. There seems to be a need to find ways of portraying this experience and this milieu so that prospective users of new programmes can relate them to their own experience:, circumstances, concerns and.preferences. The case-study is one such way, and evaluators have been proMinent ameng those who are beginning to advocate, and explore in practice, its fruitfulness 'in educational enquiry: The purpose of this brief paper is to examine the pedigree of the case-study as a research method,' to draw'attention to(featurPs of our edunational.system which pose problems for the conduct of case-study, and to suggest some guidelines for its use in education. ,
• Case-study is the examination of an instance in action. The choice of the word 'instance' is. significant in this definition, because it implies a. goal of generalisation. We might thay that case-study is that form of research where n=1, only that would be misleading, because the case-study method has outside the discourse of mathematical experimentalism that has dominated Anglo-American educational research. It is difficult to account for the neglect of the case-study as a general method of educational science, especially in view of its significant role in the history of learning theory. An adequate explanation would need to embrace the follOwing considerations: aspirations to the status of the natural sciences via the adoption of their alleged'paradigm, bureaucratic demands for actuarial data susceptible to policy manipulation, genuine scepticism about the research value of the approach, and the late arrival on the educational scene of research practitioners with relevant shills. But there is another reason' for this neglect, one which has status implications, but which poses a dilemma for the aspiring educational scientist in a curious form. It has to do with the fact that the kind of case-studies which we believe education needs have characteristics which call for a fusion of the styles of the artist and the scientist. When Freud said, "It still strikes me myself as strange that the casehistories I write should read like short stories and that as one might say, they lack the serious stamp of science", 1 he caught the unease of
2.
the researcher who, disdaining the 'safety'of numbers', discovers that. his data is most effectively expressed in a mode which is generically associated with the artist. Although this comesas a surprise to the scientist who adopts this approach, and generally a disconcerting one, it is a logical consequence of his field of vision. Case-study is the way of the artist,who achieves greatness when, through the portrayal of a single instance locked in time and circumstance, he communicates enduring truths about the human condition. For both scientist and artist, content and intent emerge in form. There have been periods in art, especially in the novel, when the artist has consciously aspired to 'scientific' generalisation. Writers of the French naturalist school, such as Zola, created characters to represent the social type, and blurred the lines between literal truth and special pleading by carefully researching the factual settings of their fictional puppets. The naturalists were part of an intellectual movement in French society which encompassed the sciences as well as the arts, a movement which had no parallel In this country. It is interesting to ponder,/therefore, the significance of one school of British television dramatists whose Preference for documentary style, accurate research, and 'representative'. heroes, owes nothing to the legacy of a native 'comic' tradition, and may constitute a take-overbid bythem for an area neglected by the social scientist. Fusion, or confusion? If we are not simply to ignore this new 'pseudo-science', perhaps we might begin by examining carefully the ,case for the case-Study, and elaborating some rules which could effectiVely discipline its use in educational research. As a method of research, the cape-study commands a respected place in the repertoire of thedry builders from a Wide range of disciplines; medicine, law, engineering, psychology and anthropology areexamples: The case can generate a theory as well as test one; instance and abstraction go hand In hand in an iterative proceas of cumulative growth. The instance may be a patient with a particular ailment, a verdict, a bridge, a chimpanzee, or A whole community, but the research aim is the same - to reveal properties of the class to which the instance belongs. "When' we read Malinowsii we get the impression that he is stating something of general importance. Yet how can this be? He is simply writing about the Trobriand Islanders. SoMehow he has so assimilated hiMself into the Trobriand situation that he is able to make the Trobriands a microcosm of the whole primitive world." 2 (Leach) .
Clearly representativeness is an important-consideration. In fields where individual variation within a'class is limited, as in medical diagnosis or in the social anthropology of non-literate peoples, case-study is widely accepted as a valid, basis of generalisation, and adopted with Confidence. PsYchopathology is an interesting area where the issue of cultural specificity continues to dog the theories Freud based on case-studies of the Viennese bourgoisie.' Nevertheless, the truth remains that, in a very important sense, we are all Freudians now.
3.
Case-study methods are rarely spelled out in advance, except in the most general of terms, and apprenticeship is theeusual means of induction into its techniques. In the social sciences, as in bank robberies, the method of attack is characteristically a opportunistic response to the observed nature of the case. The sociologist White, looking back on his research plans for the study of Cornerville, remarks: "It seems to me that the most impressive thing about them is their remoteness from the actual study I carried out". More experienced practitioners may deflect questions about methoil with Ome panache, Alan Deals a an anthropologist, writes, "In 1952, on the way to India, I asked a distinguished British anthropologist the secret of his success in doing fieldwork. His response was, "Never accept free housing, and always carry a supply of marmalade". Dut for those with less established reputations, doubt, and unease are endemic. Thus Dollard: "Many times during the conduct of the research and the arrangement of the material, I have had a bad conscience on the score of mothod Should the researcher expect to be believed if he cannot hook his findings inte.the number system end present them,in a manner conventional in the physical sciences?" (Do117d, 1937)." Dollard subsequently came under pressure from the experimentalists on methodológical grounds', but though he applied himself lengthily to psychophysical methods thereafter, he still maintained (Dollard 1957) that case-study method in the social sciences must follow a different path. "Not every nth person Can be a friend." • Problems of Case-Studying Social
Action
Despite Dollard's flight to numerical methods, it is a mistake to think that 'a,majOr distinction between experimental and case-study research has lain in the area of quantification. In fact, the case-study worker is typically much mor6 quantitative than is appreciated. Becker writes: "The observer, possessing many provisional, problems, concepts and indicators . . wishes to know which of these are worth pursuing as major foci of his study. He does this, in part, by discovering if the events that prompted their development are typical and wideSpread, and by seeing how these events are distributed among categories of people and organisational sub-units. He reaches conclusions that are essentially quantitative, using them to describe the organisation he is studying," 7 —
Much more to the point is the fact that, whereas experimental method is conceptually asocial, the most important feature of case-study in the human sciences is that it is pursued via s social process and leads to a social product. Although this is of course true of all human research, in case-study the process is significantly more intimate, the product more directly consequential for those involved, A list of the problems which the case-study worker encounters includes therefore: - Problems of the researcher becoming involved in the issues, events or situations under study: - Problems over confidentiality of data. • - Problems stemming from competition from different interest groups for access to and control over the data.
4. -
'Problems concerning publication, such as the need to preserve anonymity of subjects. Problems arising from the audience being unably to distinguish data from the researcher's interpretation of the date.
And prior to these, although linked to them, there is the problem of how to gain access to the data. The investigator of social meShanisms is seldom free to follow Dacon's dictum for Natural Sciences,. "to-put Nature to the question", or, as Collingwood put it in laying a slmlar duty upon his historical method, "to compel Nature to answer". For the observational scientist there is no 'command performance'. He must find vantage points and roles within a web of human relationships without destroying the fabric. The delicacy and subtlety of his instruments is a pre-condition of their validity, especially in those situations where consciousness Of the researcher's purposes evokes behavioural illusions designed to protect self-serving images. Educational situations are typically of this kind. Indeed, education is a field which is likely to raise in an acute form most of the problems endemic to case-study research. This being correct , we would argue that case--studies in education need to be conceived and cnnducted in new ways if they are to gain widespread acceptability. -
-
.
The Social Philosophy of Educational Case-Studies Case-study research is used occasionally in education but mostly in an 'autocratic' mode. That IS to say the problems listed in the previous section are perceived as 'technical' issues, irritating obstructions to the scientists uncompromising pursuit of new knowledge. • The study is typically conceived as a piece of 'pure' research directed to an audience of research professionals. The interpretations made are essentially the interpretations of the researcher. The responsibility for the final account is his alone. The researcher's right to control his work in these ways has not been seriously questioned. We believe it is time'to question it now. The critical issues which emerge in this reappraisal can be posed quite simply: .
-
to whose needs and interests does the research respond?
-
who owns the data (the researcher, the subject, the sponsor)?
- who has access to the data? (Who is excluded or denied)? -
what is the status of the researcher's interpretation of events, vis-a-vis the interpretations made by others? (Who decides who tells the truth?)
-
what obligations does the researcher owe to his subjects, his sponsors, his fellow professionals, others?
-
who is the research for?
Research is primarily concerned with the creation, organisation and dissemination of knowledge. COnventionally, dissemination comes last in the order: In some definitions it is omitted altsgether:". We believe that the dissemination of new knowledge ought to be a prior, 'not a post, consideration in the planning and conduct of educational research. -
Knowledge is the basis on which many forMs of power are legitimated, and, in the case of education, the medium through which power is exercised.
5.
Case-study research in education takes the researcher into a complex act of politically sensitive relationships. In a related paper (MacDonald /974) one of the authors classified evaluation studies under three ideal types on a political dimension. He called these types 'bureaucratic', 'autocratic' and 'demoeratie. The principal question which determines- this • classification is "Who controls the pursuit of new knowledge, and who has access to it?" One of the purposes behind this paper is to . endorse specifically the 'democratic' approach as particularly appropriate in case-stlidy researcher evaluation activities using case-study techniques. Democratic evaluation is described as follows: "Democratic evaluation is an information service to the community about the characteristics of an educational programme. It recognises value Pluralism and seeks to represent a range of interests in its issue formulation. The basic value is an informed citizenry, and the evaluator acts as broker in eichanges of information between differing groups. His techniques of data gathering and presentation must be accessible to non-specialist audiences. His main activity is the collection of definitions of, and reactions to, the programme. He offers confidentiality to informants and.. gives them control over his use of information. The report is non recommendatory, and the evaluator has no concept of information misuse. The evaluator engages in periodic negotiation of his relationships with sponsors and programme participants. The the range of audiences served. The . criterion of success report aspires to 'best. seller' Status. The key concepts of democratic evaluation are .sonfidentiality', 'negotiation' and 'accessibility'. The key justificatory concept is the 'right to know 1 ." 9 We feel : there is a need to develop case-study in education within this mode. Although the concepts and principles we will advance fall short of this they embody an appreach that is sharply differentiated from past or current practice, an approach which rejects monopolistic concepts of control and access. These concepts and principles also take account of significant, but neglected, features of the educational system. In this concluding section of the paper, some of these features will be specified and related to characteristics of case-study before we present our recommendations. 1. Sisnificant Features of the Educational System a) Happy alliances between theorist and practitioner in our system are rare: more often, the relationship is one of mutual mistrust punctuated by open antagonism. Detween sub-groups of practitioners also, and perhaps particularly between teachers and managers, the unity of common purpose rests on almost religious observance of territorial boundaries. Practitioners can, however, generally rely on each other for support when faced with an external enemy, such as public criticism, whereas the theorist's behaviour in such circumstances is less predictable.
b)
Partly as a consequence of this, education has a highly developed and long standing mythology which acts as a protective public image projected by its members. At all levels of the system what people think they are
doing, what they say they are doing, what they appear to others to be doing, and what in fact they are doing, may be sources of considerable discrepancy.
0.
This is generally as true of children in class as it is of teachers, head teachers and administrators. Any research which threatens to reveal these discrepancios threatens to create dissonance, both personal and political. • .
c) There is a tradition of freedom from scrutiny by outsiders in education (the inspectorial role being mainly benign and supportive). d) Educational institutions are hierarchical and competitive, for staff and for children. Dut in our country, teachers, headmasters, inspectors and administrators all have similar professional qualifications, so the hierarchy of staff is basod'oll experience and 1$ consultative. Expertise cannot be claimed on extrinsic criteria and used as a basis for authority. This characteristic of the structure, allied to mythological aspects df the culture, creates an inherent need for secrecy that is all pervasive. Case-study research may penetrate the secrecy and so threaten the carefully constructed claims which form the basis of authority. e) The educational enterprise continuously generates its own reflective languages. Since the process itself contains its own theoretical, analytical eand descriptive constructs, this creates the potential for presenting case-studies within the language of those studied. -
2. Related Aspects of Case-
d'
a) Case-studies are public docnwents alsout individuals and events. They are identifiable at' least to those involved and usually to wider audiences. They have consequences for the lives of those portrayed as well as for'the reader. b) Educational case-Studies are usually'financed by people who have, directly or indirectly, power over those studied and portrayed. • c) Case-study methods rely heavily on human instruments, about which only limited knowledge can be obtained and whose private expectations, desires and interests may bias the study in unanticipated and unacknowledged ways. Lack of rules for case-study leaves research opportunities open to both real and imagined abuse. d) Case-studies are always partial accounts, involving selection at
every stage, from choosing cases for study to sampling events and instances, and to editing and presenting material. Educational case-sttnUes are almost always conducted under constraints of time and resources and therefore reliability and validity pose eonsiderable problems. •
3. Proposed Guidelines a) It seems to us feasible to contemplate a form of educational. research that would be practice-based in a way research has not been previously. We see this as a form of research which responds actively to practitioners' definitions of situations, conceptual structures and language. Constrained by these the,researcher would act as'the representative of the various group involved, exploring their hypotheses,
7.
using their language and conceptual structures, both as starting points and as a continuous reference. The aim would not be to create alternative realities for practitioners but to find ways of encouraging them to develop insight into their existing realities, and to understand the realities of other inhabitants. c) We have to think in terms of 'condensed field work' ia order to feed back information quickly enough to fit the time-scales of participants. This cuts us off from traditions of case-study work in the social sciences and draws us closer to other traditions in journalism, documentary film-making and the novel. These traditions have not systematically addressed the issues of reliability and validity which condensed field-work will raise in acute form. Now criteria are needed. d) Proof is rarely obtainable in case-study research. Rather than setting proof es a primary goal, the case-study worker should aim to increase understanding of the variables, parameters and dynamics of the case under study. Cross-checking rather than consistency is the main strategy of validation. The case-study worker is guided in his research by the pursuit of discrepancy. It ie implicit in the notion of case-study that there ie no one true definition of the situation. In moeial situations, truth is multiple. The case-study worker is a collector of definitions. The collection is validated via a continuous process of negotiation with, those involved. Wherever appropriate the case-study should con elie expressed reactions (unedited and unglossed) of the principal characteee portrayed, to the report in its final draft form. The reliability of the stude, i.e. the probability of its findings being confirmed by replication, sees likely to be significantly enhanced where such proceduees are adopted. e) Confidentiality will become a critical aspect of procedure. Confidentiality should be accorded to informants for the term of the study and, thereafter, release of material for publication negotiated with them. The nature of case-study is such that participants can often only judge the consequences of release in retrospect and when the content of presentation is available. Offering blanket confidentiality affords the researcher faster access to relevant data and prevents the need for informants to continuously monitor what they say. This is especially critical in interview situations, for in the hands of a skilled interviewer most people are inexperienced and will reveal things they did not intend. Allowing retrospective control of editing and release of data to informants affords them some protection from the penetrative power of
the research as well as allowing the researcher to check on misinterpretation or misunderstanding on his part. This sharing of control over data with participants does mean that the researcher often has to face the fact that some of his finest data is lost, diluted or permanently consigned to the files. On the other hand his access to knowledge about what are sensitive issues to his informants may guide his research in significant and unexpected ways. f) Case-study methods lend themselves to a variety of means of presentation; written reports, audio-visual recordings, displays,and exhibitions. Generally, these presentations should be devoid of indications of praise or blame from the point of view of the researcher. They should present contingency relationships only, leaving it to those studied and other audiences, to infer cause. They should attempt to be explicit about rationale and procedure, and the principles governing the selection and presentation of content.
a.
g) The data must be accessible to the Judgement and understanding of all those Whose interests may be influenced by its contents. If this is not to remain at the lipseryice level, the. researcher may
have to explore relationships between himself and the artist.
SUMMARY We have presented a view of case-study research in education which has its primary focus on the political nature of relationships -between the researcher and his subjects, sponsors, audiences and related groups.
We have emphasised such questions as who has control over, or access to data, and under what conditions and constraints should the researcher Seek and present his findings. Our recommendations are derived from a
particular socio/philosophical stance. As we imagine and describe it, educational case-study has as yet no practitioners. We have attempted to describe a kind of research we feel ourselves working towards rather than one we have successfully accomplished. The real prize is the prospect of developing techniques and procedures which can be used by schools and ancillary agencies. A specialist research profession will always be a poor substitute for aself-monitoring educational community. .
August 1974
REFERENCES 1. Freud, S. "Standard Edition", Vol. 2, p. 160. Rogarth Press and the Institute of Psychoanalysis, 1953. 2. Leach, E R. "Rethinking Anthropology". Athlone Press, 1062. 3. Whyte, W F. "Street Corner Society", University of Chicago Press, 1055. 4. Deals, A. "Field Work in Eleven Cultures", edited by G Spindler. Holt, Rinehart and Winston, 1972. 5. Dollard, J. "Caste and Class, in a Southern Town". Harper, 1940. .
6. Dollard, J. op. cit. (1957 edition). 7. Becker, Howard S. "Sociological Work", Allen Lane, Penguin Press, 1071. O. Collingwood, R J. "The Idea of History". Clarendon Press 1946. OUP, paperback, 1961. 9. MacDonald, Larry. "Evaluation and the Control of Education", in Curriculum Evaluation: The State of the Arty Schools Council Research Studies, (in press). NOE.
Most of the points raised in this gaper are treated more fully by ROD WALKER in a longer SAFARI article entitled "The Conduct of E4ucational Case-Study: Ethics, Theory and Procedure."